in-plot data manipulation: reversible filter (masking?)

[EvaMaeRey]

I'm wondering if people had ideas about a reversible in-plot data manipulation filter method. Sometimes, I might like to filter, but then downstream recover the full data (I think).

library(ggplot2)
data_filter <- function(.keep, .by) {
  structure(list(keep_specification = rlang::enquo(.keep), 
                 by_specification = rlang::enquo(.by)), 
            class = "filterobs")
}

ggplot_add.filterobs <- function(object, plot, object_name) {
  
  new_data <- dplyr::filter(plot$data, 
                            !!object$keep_specification, 
                            .by = !!object$by_specification)
  plot$data <- new_data
  plot
  
}

library(ggplot2)

ggplot(mtcars) + 
  aes(wt) +
  geom_density() +
  aes(color = factor(cyl))

last_plot() +
data_filter(cyl != 4)

I kind of am inspired by this data_nest/data_unnest(). Which lets you look at unique observations based on a var or sets of vars. My go to would be distinct, if I weren't interested in being able to restore the data. But, I distinct drops a lot of info, while nest collapses it. Not sure that this can actually help answer the filter question.

data_nest <- function(.by) {
structure(list(by_specification = rlang::enquo(.by)),
class = "data_nestvar")

}

ggplot_add.data_nestvar <- function(object, plot, object_name) {

new_data <- tidyr::nest(plot$data,
.by = !! object$by_specification)

plot$data <- new_data
plot

}

data_unnest <- function(cols) {
structure(list(),
class = "data_unnestvar")

}

ggplot_add.data_unnestvar <- function(object, plot, object_name) {

new_data <- tidyr::unnest(plot$data, cols = "data")

plot$data <- new_data

plot

}

library(ggplot2)
diamonds |>
ggplot() +
aes(x = cut) +
geom_bar()

last_plot() +
data_nest(.by = cut)

last_plot() +
data_unnest()

[EvaMaeRey]

Trying this out currently. 😬 Saving unfiltered data in new plot list item unfiltered_data for filtering moves:

plot$unfiltered_data <- plot$unfiltered_data %||% plot$data

Then an unfiltering function restores initial data input.

plot$data <- plot$unfiltered_data